Insulin and the early bovine embryo

Insulin and the early bovine embryo

Influences on in vitro development, gene expression, and morphology

Denise Laskowski

Faculty of Veterinary Medicine and Animal Science Department of Clinical Sciences

Uppsala

Doctoral thesis

Swedish University of Agricultural Sciences

Uppsala 2017

Acta Universitatis agriculturae Sueciae

2017:65

ISSN 1652-6880

ISBN (print version) 978-91-7760-020-6 ISBN (electronic version) 978-91-7760-021-3

© 2017 Denise Laskowski, Uppsala Print: SLU Service/Repro, Uppsala 2017 Cover: Drawing by Britta Laskowski

(photo: Denise Laskowski)

Metabolic imbalance is a problem in the dairy industry because the metabolic demands of increased milk production can lead to decreased fertility, and more knowledge about improving the management and physical conditions of the cow (the links between fertility, nutrition, milking, and dry period) is needed. Insulin is an important hormone regulating the energy balance in the body, and insulin concentrations change in situations of energy deficiency or excess, both of which are known to decrease fertility in cows as well as in humans. Hyperinsulinemia is associated with decreased fertility by impairing the developmental potential of embryos, but the underlying reasons for this remain unclear. Our aim was to investigate insulin-induced changes on development, morphology and molecular signature in bovine blastocysts on Day 8 (BC8).

An in vitro model was used and morphology and gene expression were analysed by combining confocal microscopy and microarray-based transcriptome studies. Blastocysts were produced in vitro according to standard methods using oocytes that were supplemented with three different insulin levels (INS10 =10 —g/ml; INS0.1= 0.1 —g/ml; INS0=control) during maturation. The transcriptome profile of BC8 was obtained and embryo quality grades, developmental stages and morphologies were further assessed in terms of F- actin, DNA, and active mitochondria. Significant differences were observed in developmental rates and morphology after insulin exposure. The observed changes were reflected by increased expression of genes involved in cell division and structure, mitochondrial activation, lipid metabolism, and oxidative stress.

Combining all of the results, it was shown that elevated insulin impairs the developmental potential of the embryo. This work contributes to new knowledge about the molecular background of embryos developing under metabolic stress conditions such as hyperinsulinemia. Moreover, the studies are of comparative value for humans where impaired fertility is often related to metabolic disorders.

Keywords: Oocyte maturation, bovine blastocyst, transcriptome, metabolic syndrome, diabetes, hyperinsulinemia, embryo in vitro production, embryo morphology, lipids Author’s address: Denise Laskowski, SLU, Department of Clinical Sciences, P.O. Box 7054, SE-750 07 Uppsala, Sweden. E-mail: denise.laskowski@slu.se

Insulin and the early bovine embryo - influences on in vitro development, gene expression, and morphology

Abstract

“Und jedem Anfang wohnt ein Zauber inne, Der uns beschützt und der uns hilft, zu leben.”

”A magic dwells in each beginning, protecting us, telling us how to live.”

Hermann Hesse (“Stufen”)

To my parents Gudrun and Olaf, and my

beloved family and friends

List of publications 9 Additional publications related to the thesis 11

List of tables 13

List of figures 14

Abbreviations 15

1 Introduction 19

1.1 Bovine reproduction and metabolic disorders 19

1.2 Insulin 20

1.2.1 Insulin receptor signalling and functions 21 1.2.2 The role of insulin in the metabolism of ruminants 23 1.2.3 Insulin and metabolic imbalance 24

1.2.4 Insulin resistance 26

1.2.5 Insulin and insulin-like growth factors 29

1.2.6 Insulin and fertility 30

1.2.7 Insulin and in vitro embryo production 32

1.2.8 Comparative aspects 33

1.3 Early embryonic development 34

1.3.1 Oocyte maturation 34

1.3.2 Development until the blastocyst stage 35

1.3.3 Embryo morphology 36

1.3.4 Oocyte and embryo metabolism 37

1.3.5 In vitro produced embryos 40

1.3.6 Comparative aspects between human and bovine embryo development 40

1.4 Gene expression in early embryos 41

1.4.1 Transcriptome 41

1.4.2 Metabolic programming and epigenetics 43

2 Aims of the thesis 45

Contents

3 Materials and methods 47

3.1 Experimental design 47

3.2 Oocyte origin and in vitro production of embryos 48

3.2.1 Oocyte collection 48

3.2.2 In vitro maturation and insulin supplementation 49

3.2.3 In vitro fertilization 49

3.2.4 In vitro culture 49

3.2.5 Measurement of insulin concentrations by ELISA 50

3.3 Embryo morphology 50

3.3.1 Developmental rates, stage, and grade 50

3.3.2 Embryo staining 51

3.3.3 Morphological evaluation of embryos 51 3.4 Microarray-based transcriptome studies of embryos 54

3.4.1 Embryo pooling and freezing 54

3.4.2 DNA/RNA extraction and RNA amplification and labelling 55 3.4.3 Hybridization of aRNA to Agilent oligo microarray slides 55 3.4.4 Microarray data analysis of the transcriptome 55 3.4.5 RT-qPCR validation of candidate genes 56 3.5 Lipid profile by desorption electrospray ionization mass spectroscopy

(DESI-MS) 56

3.6 Statistical analysis 57

3.6.1 Developmental rates, morphological analysis, and cell counts 57

3.6.2 Microarray 58

3.6.3 RT-qPCR 58

3.6.4 DESI-MS 58

4 Main results and discussions 61 4.1 Method validation study of in vitro insulin exposure (Paper I) 61 4.2 Influence of insulin on development rates (Paper II and III) 64 4.3 Insulin and embryo morphology (Paper II and III) 65 4.4 Insulin and embryo gene expression (Paper II, III, IV) 68

4.4.1 General transcriptome patterns of the embryo and the most relevant pathways activated by insulin during oocyte maturation

(Paper III) 68

4.4.2 Signatures of an impact of insulin on embryo lipid metabolism

(Paper IV) 71

4.5 Insulin and embryo lipid profile (Paper IV) 72

5 Concluding remarks 75

6 Future perspectives 77 References 81

Popular science summary 107

Populärvetenskaplig sammanfattning 109 Acknowledgements 111

This thesis is based on the work contained in the following papers, referred to by Roman numerals in the text:

I Laskowski, Denise; Sjunnesson, Ylva; Gustafsson, Hans; Humblot, Patrice; Andersson, Göran; and Båge Renée (2016). Insulin concentrations used in in vitro embryo production systems - a pilot study on insulin stability with an emphasis on concentrations measured in vivo. Acta Veterinaria Scandinavica 58 (Suppl.1): 66.

II Laskowski, Denise; Båge, Renée; Humblot, Patrice; Andersson, Göran;

Sirard, Marc-André and Sjunnesson, Ylva (2017). Insulin during in vitro oocyte maturation has an impact on development, mitochondria, and cytoskeleton in bovine Day 8 blastocysts. Theriogenology 101, 15–25.

III Laskowski, Denise; Sjunnesson, Ylva; Humblot, Patrice; Sirard, Marc- André; Andersson, Göran; Gustafsson, Hans and Båge, Renée (2016). In vitro bovine oocyte maturation changes blastocyst gene expression and developmental potential. Reproduction, Fertility and Development 29, 876-889.

IV Laskowski, Denise; Andersson, Göran; Humblot, Patrice; Sirard,

Marc-André; Sjunnesson, Ylva; Ferreira, Christina; Pirro, Valentina; Båge, Renée (2017). Lipid profile of bovine blastocysts exposed to insulin during in vitro oocyte maturation. Reproduction, Fertility and Development (submitted manuscript).

Papers I-III are reproduced with the permission of the publishers.

List of publications

I Literature review, initial planning of the pilot study (including choice of measurement methods), the IVF laboratory work, and preparation of samples. Main responsibility for writing the manuscript.

II Performed most of the laboratory work, microscopy imaging and data analysis and interpretation together with supervisors. Main responsibility for writing the manuscript.

III Planning and performance of the laboratory work together with supervisors and collaborators, performed the IVF work, performed molecular biology methods under supervision, took part in data analysis and performed data interpretation. Main responsibility for writing the manuscript together with collaborators and supervisors.

IV Took major part in planning of the lipid study, establishing contact with collaborators, and developing methods. Interpretation of gene analysis data together with supervisors and collaborators, main responsibility for writing the manuscript.

The contribution of Denise Laskowski to the papers included in this thesis was as follows:

¾ Laskowski, Denise; Sjunnesson, Ylva; Humblot, Patrice; Andersson, Göran;

Gustafsson, Hans; and Båge, Renée (2016). The functional role of insulin in fertility and embryonic development - What can we learn from the bovine model? Theriogenology, 86(1), 457-464.

¾ Laskowski, Denise; Humblot, Patrice; Sirard, Marc-André; Sjunnesson, Ylva; Jhamat, Naveed; Båge; Renée; Andersson, Göran (2017). Elevated insulin changes DNA methylation pattern in bovine in vitro blastocysts.

Submitted to Reproduction.

Additional publications related to the thesis

Table 1. Summary of the morphological categories and characteristics used to

asses actin and mitochondria. 54

Table 2. Number of oocytes, cleavage rate, and blastocyst rate on Day 8 (BC8). 64 Table 3. Number of nuclei (TN= total number) and actin and mitochondria

categories. 66 Table 4. Influence of 10 —g/ml (IN10) and 0.1 —g/ml insulin (INS0.1) during

oocyte maturation on gene expression. 70 Table 5. Cholesterol-related genes with significant p-value (p<0.05) and fold

change differences in INS10 (a) and INS0.1 (b) compared to INS0, sorted according to their function in lipid metabolism. 72

List of tables

Figure 1. Insulin signalling pathway (described in detail in text). 22 Figure 2. Involvement of ROS in mitochondrial dysfunction and insulin

resistance. 26 Figure 3. Insulin signalling during insulin resistance. 29 Figure 4. Proposed model of the metabolic interactions and activity of cumulus

cells and the oocyte. 39

Figure 5. Experimental design and overview of studies I to IV. 48 Figure 6 a and b: Staining of nuclei with Hoechst dye and cell counts in BC8. 52 Figure 7. Actin and mitochondria staining and categories. Examples of actin

categories in F-actin (Alexa Fluor 488 Phalloidin) stained BC8 53 Figure 8. Binding of circulating insulin and IGFs to target cells. 62 Figure 9. Interactions of insulin regulated pathways in the lipid metabolism. 74

List of figures

ACAA1 Acetyl-CoA acyltransferase 1 ACP5 Acid phosphatase 5

ADIPOR2 Adiponectin receptor 2 AGC Automatic gain control Akt/PKB Protein kinase B

APLP2 Amyloid precursor-like protein 2 APOA1 Apolipoprotein A 1

aRNA Anti-sense ribonucleic acid ART Assisted reproductive technologies

BC8 Blastocyst day 8

BSA Bovine serum albumin

CAM Camera image of the epifluorescence microscope

CC Cumulus cells

COC Cumulus-oocyte complexes COMT Catechol-O-methyltransferase

CYP11A1 Cytochrome P450 family 11 subfamily A member 1 DESI MS Desorption electrospray ionization mass spectrometry DET Differentially expressed transcript

DG Diacylglycerol

DHCR7 7-Dehydrocholesterol reductase DMR Differentially methylated region EGA Embryonic genome activation EHD1 EH domain containing 1

ELISA Enzyme-linked immunosorbent assay

ER Endoplasmic reticulum

FA Fatty acid

FADS2 Fatty acid desaturase 2

Abbreviations

FSH Follicle stimulating hormone

gDNA Genomic DNA

GLUT-4 Glucose transporter type 4 GnRH Gonadotropin releasing hormone GRB2 Growth factor receptor-bound protein 2 H2O2 Hydrogen peroxide

HDL High density lipoprotein

HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid HMGCR 3-Hydroxy-3-methylglutaryl-CoA reductase HSPA1A/B Heat shock protein family A (Hsp70) Member 1A

ICM Inner cell mass

IETS International Embryo Technology Society IGF Insulin-like growth factor

IGF2R Insulin-like growth factor 2 receptor INSIG1 Insulin induced gene 1

IRS Insulin receptor substrate IVF In vitro fertilization IVP In vitro production

LAMP1 Lysosomal-associated membrane protein 1 LCFA-CoA Long-chain acyl-coenzyme A

LDLR Low density lipoprotein receptor

LH Luteinizing hormone

MAP Mitogen-activated protein MAPK Mitogen-activated protein kinase MEK Mitogen-activated protein kinase kinase

mRNA Messenger RNA

mSOF Modified synthetic oviductal fluid

mTALP Modified Tyrode's albumin lactate pyruvate MVD Mevalonate diphosphate decarboxylase m/z Mass number/ion charge number NEB Negative energy balance

NEFA Non esterified fatty acids

NR1H2 Nuclear receptor subfamily 1 group H member 2 NR3C1 Nuclear receptor subfamily 3 group C member 1 PBS Phosphate buffered saline

PEB Positive energy balance

PIK3 Phosphatidylinositol-4,5-bisphosphate 3-kinase PIP2 Phosphatidylinositol 4,5-bisphosphate

PIP3 Phosphatidylinositol (3,4,5)-trisphosphate PNLIP Pancreatic lipase

PPARĮ Peroxisome proliferator-activated receptor alpha PTPases Tyrosine-specific protein phosphatases

PVA Polyvinyl alcohol

Raf Rapidly accelerated fibrosarcoma Ras Ras protein family

RIN RNA integrity number ROS Reactive oxygen species

RT Room temperature

RT-qPCR Reverse transcribed quantitative PCR

SCAP Sterol regulatory element-binding protein cleavage- activating protein

SH2/3 Src Homology 2/3 domain

SH2/ShC Src homology 2 domain-containing

SoS Son of Sevenless

SREBP Sterol regulatory element binding protein

TAG Triacylglycerid

TCM Tissue culture medium TNF Tumour necrosis factor

VIM Vimentin

W/V Weight per volume

1.1 Bovine reproduction and metabolic disorders

Reproductive performance is a key factor for cost efficiency in the dairy industry and cow reproduction and early embryonic development are well-studied topics.

The oocyte matures in the follicle surrounded by follicular fluid, and this process is influenced by metabolic and nutritional changes in the blood (Leroy et al., 2004, 2011; Aardema et al., 2013). The concentrations of these metabolites might differ significantly depending on the maternal nutritional state. The oocyte appears to be very sensitive to metabolic stress, which might decrease its developmental potential. Moreover, there is even a risk that the offspring of an obese mother later in life will suffer from metabolic diseases because metabolic programming occurs early in life through epigenetic changes that can remain throughout the entire life (Heerwagen et al., 2010).

Follicular growth and oocyte maturation takes 8 -12 weeks in the dairy cow (Beam & Butler, 1997, 1999). This means that the oocyte in the growing follicle that will be subjected to the first inseminations after parturition undergoes several metabolically challenging periods until final maturation. Often, dairy cows are over-conditioned in the dry period (Rukkwamsuk et al., 1999), which leads to elevated circulating insulin levels and impaired insulin sensitivity (Locher et al., 2015). After calving, the metabolic profile of dairy cows changes due to the increasing energy demands for milk production (Fleischer et al., 2001). Most dairy cows go through a period of negative energy balance while insulin levels are low (Butler et al., 2004), and non-esterified fatty acid (NEFA) concentrations are elevated due to lipid mobilization (Bossaert et al., 2008). The dairy cow suffers from ketosis, a state during which ȕ-hydroxybutyric acid levels are increased, a sign for severe energy deficiency leading to secondary diseases due to immune system deficiencies (Ingvartsen, 2006). Circulating insulin levels

1 Introduction

increase when positive energy balance is restored (Gong et al., 2002; Humblot et al., 2008), and this has positive effects on reproductive functions. This switch from energy excess to energy deficiency might negatively affect the developmental competence of the oocyte because the oocyte metabolism and gene expression have to be well balanced to allow for optimal oocyte development during that time period (Leroy et al., 2011). It is known that subfertile dairy heifers rapidly gain weight and become overconditioned after a period of repeated, unsuccessful inseminations (Gustafsson, 1985). These heifers are characterized by hormonal aberrations and by impaired oocyte quality and reduced early embryonic developmental competence (Båge, 2002;

Båge et al., 2003; Awasthi et al., 2010). If fat heifers finally manage to conceive, they will be at high risk of dystocia at parturition along with retained placenta and metritis, which in turn will increase the risk for fertility problems in the insemination period.

It has become evident that both increases and decreases in circulating insulin levels have a potential role in the health of the embryo developing under such conditions.

Moreover, similarities between the human and bovine species provide a fruitful research field because new insights in the relation between metabolic and reproductive disorders could help to improve fertility in both species. The suitability of a bovine model for human embryonic development can be explained by analogies in ovarian reserve, follicular dynamics and embryonic metabolism (Ménézo & Hérubel, 2002; Campbell et al., 2003a).

1.2 Insulin

The hormone insulin was discovered in 1921 by Banting and Macleod and this marked a major breakthrough in medicine (Quianzon & Cheikh, 2012). Long before that, it was already hypothesized that there must be a substance secreted by the pancreas controlling carbohydrate metabolism (Bliss, 1993). Macleod isolated insulin from cow pancreases, and the first diabetic patient was successfully treated in 1922. Since then, insulin has been in focus for understanding the pathophysiological mechanisms of diabetes and related diseases, and the successful therapy of patients suffering from diabetes was honoured with the Nobel Prize in 1922.

Insulin is a peptide hormone produced by the ȕ-cells in the islets of Langerhans (Sonksen & Sonksen, 2000), and it functions as a key metabolic regulator of energy homeostasis in the body. Insulin acts on multiple levels of lipid and glucose metabolism (Saltiel & Kahn, 2001), and its metabolic effects include stimulation of DNA synthesis, protein synthesis, transmembrane

transport, and lipogenesis (Kahn & White, 1988). Insulin is proteolytically converted from its precursor polypeptides in several steps, and the final activation step occurs by cleaving C-peptide from proinsulin (Steiner, 2008).

The biologically active insulin molecule consists of 51 amino acids forming the A and B chain that are connected via three disulphide bridges (Brange &

Langkjœr, 1993). Insulin is highly evolutionarily conserved in its structure among species; for example, bovine and human insulin only differ in three amino acid residues (Smith, 1966).

The interest in understanding the action and regulation of insulin has not decreased in recent years, and research on insulin is still of high relevance because diabetes and diabetes-related diseases are increasing worldwide (Seidell, 2000). Knowing about the importance and history of insulin explains our aim to contribute with our research to add one more piece to the puzzle in better understanding insulin-related pathways during early embryonic development.

1.2.1 Insulin receptor signalling and functions

Insulin action is transmitted through its binding with different affinities to both the insulin receptor and the related insulin-like growth factor (IGF)-1 receptor (Rechler & Nissley, 1985). The insulin receptor belongs to the family of ligand- activated receptors with tyrosine kinase activity, and it contains transmembrane signalling domains with two alpha and two beta chains (Lee & Pilch, 1994).

Both insulin and IGF receptors auto-phosphorylate tyrosine residues on the receptor upon ligand binding, and their biological response is transmitted by phosphorylation of intracellular proteins that activate different signalling cascades (Sibley et al., 1987; Häring, 1991; Tsakiridis et al., 1999).

Two main intracellular pathways lead to either metabolic or mitogenic post- receptor actions through mitogen activated protein kinase (MAPK) or phosphatidyl-inositol-3-kinase (PI3K) (Shepherd et al., 1998; Figure 1, schematic illustration). In brief, insulin binding activates the insulin receptor tyrosine kinase to phosphorylate insulin receptor substrates (IRSs). IRS proteins are recruited to the receptor and phosphorylated on tyrosine residues, leading to binding sites for molecules with a Src-homology 2 (SH2) domain such as PI3K (Myers et al., 1992), Shc –proteins, and growth factor receptor-bound protein 2 (GRB2) (Pelicci et al., 1992; Sasaoka et al., 1994; Sun et al., 1997; Boucher et al., 2014). Activated PI3K generates phosphatidylinositol (3,4,5)-trisphosphate (PIP3) from phosphatidylinositol 4,5-bisphosphate (PIP2), and this second messenger leads to an activation cascade of different protein kinases, especially protein kinase B (PKB/Akt), a serine kinase. PKB is involved in the translocation

of glucose transporter type 4 (GLUT-4) vesicles to the cell membrane, the most well-studied insulin action allowing glucose uptake. PKB also leads to increased protein synthesis (Pessin & Saltiel, 2000; Bevan, 2001; Lizcano & Alessi, 2002).

The signal transduction proteins leading to mitogenic functions are GRB2, a protein with SH2 and SH3 domains, and Shc (Pelicci et al., 1992; Sasaoka et al., 1994; Chow et al., 1998). Shc proteins are able to bind GRB2 after tyrosine phosphorylation which leads to activation of Ras and, the beginning of a phosphorylation cascade and the consequent activation of MAPKs (Giorgetti et al., 1994). MAPKs are proteins involved in the regulation of cell growth, division, and apoptosis and are able to phosphorylate nuclear proteins and several protein kinases that interact with transcription factors and thus influence gene expression (Sturgill et al., 1988; Bevan, 2001; Hilger et al., 2002).

Figure 1. Insulin signalling pathway (described in detail in text).

Akt/PKB = Protein Kinase B, GRB-2 = Growth factor receptor-bound protein 2, IRS = Insulin receptor substrate, MAPK = mitogen-activated protein kinase, MEK = Mitogen-activated protein kinase kinase, P = Tyrosine phosphorylated state, PI3K = Phosphatidylinositol-4,5-bisphosphate 3-kinase, PIP2 = Phosphatidylinositol 4,5- bisphosphate, PIP3 = Phosphatidylinositol (3,4,5)-trisphosphate, Raf = Rapidly accelerated fibrosarcoma, Ras = Ras protein, SH2 = Src homology 2 domain-containing, ShC = ShC protein, SoS = son of sevenless, *activated

1.2.2 The role of insulin in the metabolism of ruminants

The different functions of insulin are tissue-specific, and the induced post- receptor mechanisms and responses differ in liver, adipose tissue, and skeletal muscle (Zachut et al., 2013). In contrast to humans, the glucose metabolism of ruminants is not entirely dependent of insulin, and only a limited amount of the available glucose derives from direct intestinal absorption (Aschenbach et al., 2010). In general, peripheral glucose concentrations are lower and insulin responses are weaker in ruminants (Kaske et al., 2001). Even if the same mechanisms – including insulin-induced translocation of the glucose transporter GLUT4 towards the cell membrane to facilitate the glucose uptake in the cell – exist, lower GLUT4 and IRS-1 availability lead to a weaker response of glucose metabolism to insulin stimuli (Sasaki, 2002). Thus, only a small amount of glucose is used for lipogenesis in the adipose tissue, which uses acetate as a preferred substrate for lipid synthesis (Hanson & Ballard, 1967).

The glucose uptake in the mammary gland can be up to 92% of the total glucose consumption (Rose et al., 1997). In the mammary gland, glucose uptake is independent of insulin signalling, and GLUT4 is not detectable (Komatsu et al., 2005), and this is further evidence for the tissue specific characteristics of insulin and glucose metabolism.

Different from humans, the glucose requirements of ruminants must always – not just during fasting – be covered by gluconeogenesis from non- carbohydrate sources. The most abundant substrates derive from microbial activity in the rumen in form of short fatty acids (Brockman, 1978; Brockman &

Laarveld, 1986) that are synthesized to glucose in the intermediate metabolism.

The regulation of catabolic and anabolic actions is transmitted through glucagon and insulin where the major glucogenic substrates are propionate, lactate, pyruvate, amino acids and glycerol (Bergman, 1973). Propionate and other short-chained fatty acids and lactate have their origin in microbial fermentation.

Lactate can originate from anaerobic glucose oxidation while glycerol together with NEFAs originate from fat mobilization (De Koster & Opsomer, 2013).

Insulin stimulates glucose uptake in muscle cells and adipose tissue, while glucagon has its primary target in the liver to stimulate gluconeogenesis (Brockman, 1978). The insulin response in different tissues varies through insulin receptor and GLUT4 availability, which changes during different periods of lactation. One explanation for this regulatory system is that glucose demands vary significantly depending on milk production and can increase up to four fold from late pregnancy to early lactation (Bell, 1995). The metabolism needs to adapt to the varying requirements by increasing expression of gluconeogenic enzymes in the liver (Graber et al., 2010). At the same time, insulin stimulation is reduced (due to low circulating insulin levels postpartum or due to peripheral

insulin resistance) which leads to lipid mobilization and thus higher availability of substrates important for gluconeogenesis.

At the hepatic level, elevated insulin decreases glycogenolysis, i.e. the conversion of glycogen to glucose-6-phosphate, and decreases lipolysis, while it increases glycogen synthesis (Donkin & Armentano, 1995). This is due to the fact that the energy substrate availability is sufficient in the presence of high insulin concentrations, and this is a condition in which storage of energy would be favourable (McDowell, 1983). Some studies report a suppressive effect of insulin on gluconeogenesis (Donkin & Armentano, 1995; De Koster and Opsomer, 2013), while other authors state that gluconeogenesis from propionate is less regulated by insulin but mainly dependent on feed intake (Brockman, 1978; Danfær et al., 1995). This makes sense because in ruminants, glucose has to be synthesised from its precursors, even in situations with high substrate availability.

1.2.3 Insulin and metabolic imbalance

Circulating levels of insulin serve as an indicator of energy balance in the body.

Situations with short or long term hyper- or hypoinsulinemia exist in cattle as well as in other mammals including humans (Butler et al., 2003; Eckel et al., 2005).

In humans, the incidence of obesity, metabolic syndrome, and type 2 diabetes is increasing, all of which are examples of hyperinsulinemic conditions associated with positive energy balance (PEB)(Seidell, 2000). In the dairy cow, PEB is often maintained during the dry period when the cow is not lactating (Holtenius et al., 2003), implying a risk of developing obesity. The change from PEB to negative energy balance (NEB) is especially dramatic in cows in good body condition, and many cows fail to adapt to the transition from an anabolic situation during the dry period to a catabolic postpartum condition (Roche et al., 2000). The failure of adapting to the increasing energy demands for lactation leads to ketosis. It is possible to distinguish two different types of ketosis, e.g.

using existing differences regarding the insulin levels measured in blood (Herdt, 2000; Oetzel, 2007). In brief, ketosis 1 or “underfeeding ketosis” often occurs 3 -6 weeks after calving at the peak of lactation and in thin animals, while ketosis 2 is present in fat cows and involves liver pathologies due to fat accumulation.

Ketosis 2 occurs one or two weeks after calving, with an initial depression in food intake followed by fat mobilization, and it has a poorer prognosis than ketosis 1. This “fat cow syndrome” type of ketosis has similarities to type 2 diabetes in humans because the glucose and insulin levels initially are high and insulin sensitivity low, and this leads to the development of insulin resistance

(Oetzel, 2007). Interestingly, different cow breeds show variation concerning their sensitivity to metabolic perturbations and in how quickly PEB can be restored (O’Hara et al., 2016; Ntallaris et al., 2017), and there are strategies to change feeding regimes to avoid the negative effect of low insulin and glucose levels (Holtenius et al., 1996; Gong et al., 2002).

In conclusion, insulin-raising feeding regimes have a beneficial effect on cow health, while fat supplementation should be avoided because it decreases feed intake and is unfavourable for metabolic functions because lipid metabolism is already overloaded by lipids from body fat mobilization. Insulin is an important regulator of the growth hormone – insulin-like growth factor (GH-IGF) axis and thus has a role in lipid mobilization. It stimulates hepatic expression of GH receptors and IGF-1, which leads to high peripheral IGF-1 concentrations, while it inhibits GH receptors and IGF-1 in adipocytes (Butler et al., 2003). High circulating levels of IGF-1 and insulin reduce circulating GH concentrations which reduces lipolytic mobilization of body energy reserves. This is the reason why it is beneficial to feed diets that increase circulating insulin levels after calving, thus reducing the negative cascade of metabolic disturbances linked to lipid breakdown. During situations of NEB (Lucy, 2006; Kawashima et al., 2007), such as in early lactation of the dairy cow or other conditions when energy intake is lower than energy requirements for maintaining body functions and all physical activity, the opposite occurs and increasing GH concentrations lead to lipid mobilization while insulin is low. Metabolic stress conditions in the form of energy excess or deficiency have shared metabolic characteristics. The elevated levels of lipids in the blood can be dietary induced or can develop through fat mobilization (Leroy et al., 2015). The over-conditioned cow in the dry period has a greater risk of suffering from NEB and secondary diseases after parturition, and this is why the extreme and abrupt change from PEB to NEB is reported to be a high stress factor that the individual has to cope with (Rukkwamsuk et al., 1999). Insulin functions in carbohydrate and lipid metabolism are well studied and described (Saltiel & Kahn, 2001), but the consequences of elevated insulin exposure during early development might differ from its effects on differentiated tissues such as the liver, adipose tissue and intestine. Oxidative stress and mitochondrial damage have been described in the context of metabolic stress (Ceriello & Motz, 2004; Furukawa et al., 2004;

Roberts & Sindhu, 2009; Wu et al., 2015) and might thus also be one explanation for the detrimental effect of changes related to elevated insulin concentrations during oocyte maturation (Figure 2).

Figure 2. Involvement of ROS in mitochondrial dysfunction and insulin resistance.

ROS = Reactive oxygen species, IRS = Insulin receptor substrate, H2O2 = Hydrogen peroxide, PTPase = Tyrosine-specific protein phosphatases, DG = Diacylglycerol, LCFA-CoA = Long-chain acyl-CoA, TFN = tumour necrosis factor, NEFA = Non-esterified fatty acids.

1.2.4 Insulin resistance

Insulin resistance is the state when physiological concentrations of insulin no longer lead to normal responses in the target cells (Kahn, 1978).

It is important to be aware that insulin resistance can be tissue-specific and a transient condition in the dairy cow (Kahn, 1978; Muniyappa et al., 2008). The most well-described form of “classical“ insulin resistance is the reduced sensitivity of muscle and adipose tissue which can occur during pregnancy, obesity, and type 2 diabetes in humans (Häring, 1991; Pessin & Saltiel, 2000).

In the dairy cow, insulin resistance is most common during late pregnancy and early lactation, most probably linked to hormonal and metabolic changes during lactation where glucose is the favourable substrate for the udder (Bell &

Bauman, 1997). Moreover, this natural regulation mechanism might coincide with a massive increase in NEFAs which are thought to reduce insulin sensitivity and furthermore contribute to insulin resistance in the peripartum period (Oikawa & Oetzel, 2006; Pires et al., 2007) by IRS-1 phosphorylation (Le Marchand-Brustel et al., 2003).

The insulin receptor expression pattern is regulated in a similar way in all insulin sensitive tissue types, and insulin exposure leads to receptor downregulation that is regenerated within four hours after the end of insulin exposure (Marshall & Olefsky, 1981). This might be one of the pathophysiological mechanisms of insulin resistance in chronic hyperinsulinemia where the physiological receptor regeneration no longer occurs (Wigand & Blackard, 1979). The mechanisms that are involved in deactivating insulin action are very important for controlling metabolic pathways. Here, phosphatases are described as the main players to dephosphorylate the tyrosine on the activated receptor and IRS (Goldstein et al., 1998).

Other factors that are known to participate in decreased insulin sensitivity are increased circulating NEFAs, hormones, proinflammatory cytokines, and reactive oxygen species (ROS), all of which are present in obese conditions and are involved in detrimental actions on post-receptor cascades (Reaven et al., 1988; Boden, 1997; Houstis et al., 2006; Shoelson et al., 2006). One mechanism of insulin resistance at the post-receptor level is serine/threonine phosphorylation of insulin IRS-1 and 2 instead of tyrosine phosphorylation, which leads to reduced activation of PI3K and the subsequent glucose transport activation steps (Figure 3 and Czech & Corvera, 1999; Farese, 2002).

Accumulation of lipids – both if originating from nutritional sources (PEB, overfeeding, obesity) or from body-fat breakdown (NEB) – has toxic effects (“lipotoxicity”). Interestingly, overconditioned subfertile heifers not only have an excess of subcutaneous fat, but they also have a parallel, detrimental accumulation of fat in their oocytes (Awasthi et al., 2010). Lipotoxicity can explain some mechanisms responsible for insulin resistance, and increased fatty acid accumulation in the cell can directly reduce insulin sensitivity. In addition to this direct effect, increased lipid influx can also indirectly contribute to insulin resistance by inducing oxidative stress during fatty acid breakdown.

Mitochondrial oxidative phosphorylation produces ROS that have to be neutralised. Intracellular fatty acid metabolites like diacylglycerol (DAG) can directly activate serine/threonine kinases that phosphorylate IRS-1 and 2 with the consequence of decreased PI3K activation and impaired downstream signalling of the insulin receptor (Randle et al., 1963; Sesti, 2006; Boucher et al., 2014).

Moreover, high energy intake implies an increased production of ROS leading to activation of various cellular stress-response pathways, which can interfere with the physiological cellular signalling pathways (Bloch-Damti &

Bashan, 2005). If fatty acids accumulate to a large extent, anti-oxidative molecules and pathways cannot be regenerated in time, which leads to

accumulation of ROS that emerge when excessive NADH, an electron donor in Acetyl-CoA- synthesis from glucose or fat, cannot be regenerated and single electrons are transferred to oxygen (Maechler et al., 1999). The development of insulin resistance might thus represent a compensatory mechanism protecting the cell against further insulin-stimulated energy substrate uptake in an attempt to reduce oxidative stress (Ceriello & Motz, 2004). Increased circulating ROS have further detrimental effects leading to inflammatory responses, mitochondrial dysfunction, and insulin resistance (Figure 2).

In humans, decreased insulin sensitivity that finally leads to insulin resistance can often be diagnosed in overweight or obese persons (DeFronzo & Ferrannini, 1991). Here, long-term overexposure to insulin due to excess energy intake (hyperinsulinemia) is detrimental for the functioning of physiological insulin signalling and can, as a consequence, lead to the manifestation of type 2 diabetes with chronic hyperinsulinemia but a lack of response to insulin in the target cells (Kahn et al., 2006).

In the dairy cow, insulin signalling is different from humans in some aspects due to changes in insulin sensitivity linked to the lactation period. During late pregnancy and continuing postpartum, the cow develops insulin resistance in adipose tissue and muscle (Bell, 1995). This is explained by the excessive energy requirements during early lactation where most glucose goes to the mammary gland where glucose uptake is mostly regulated through GLUT1, and this form of glucose transport is insulin independent (Komatsu et al., 2005, Rose et al., 1997).

In contrast to insulin resistance in over-conditioned cows or during obesity – where hyperinsulinemic conditions appear – circulating insulin and glucose concentrations are low after parturition (De Koster & Opsomer, 2012).

This adverse condition where hypoinsulinemia and insulin resistance appear simultaneously highlights the complexity of different mechanisms involved in insulin signalling and its pathophysiological mechanisms. The beneficial or detrimental effect of elevated insulin concentrations always needs to be put in the context of the nutritional situation, and this explains why both high and low insulin levels can imply stress for the individual.

Figure 3. Insulin signalling during insulin resistance.

Akt/PKB = Protein kinase B, ER = Endoplasmic reticulum, GRB2 = Growth factor receptor-bound protein 2, IRS = Insulin receptor substrate, MAPK = Mitogen-activated protein kinase, MEK = Mitogen-activated protein kinase kinase, NEFA = Non-esterified fatty acids, P = Tyrosine phosphorylated state, PI3K = Phosphatidylinositol-4,5-bisphosphate 3-kinase, PIP2 = Phosphatidylinositol 4,5-bisphosphate, PIP3 = Phosphatidylinositol (3,4,5)-trisphosphate, Raf = Rapidly accelerated fibrosarcoma, Ras = Ras protein, ROS = Reactive oxygen species, SH2 = Src homology 2 domain-containing, Shc = Shc protein, SoS = son of sevenless,

*activated

1.2.5 Insulin and insulin-like growth factors

Insulin-like growth factors (IGFs) are mitogen polypeptides, so-called survival factors that have structural similarities and functional interactions with insulin and insulin signalling pathways. Both IGF-1 and 2 have important roles both in pre- and postnatal growth (DeChiara et al., 1991; Baker et al., 1993). IGF-1 is expressed at low levels in the embryo and thus is seen as more important for postnatal development and growth control than during the embryonic and foetal period. However, mice nulliparous for IGF-1 have been shown to have more than 60% decreased birth weight compared to wild type mice (Powell-Braxton et al., 1993).

IGF-2 is maternally imprinted and is mainly involved in prenatal growth through IGF-1 and insulin receptor. This shows that the insulin receptor, besides its important role in insulin signalling by transmitting metabolic effects, acts on growth control during embryogenesis (Louvi et al., 1997).

In the bovine, transcripts of IGF-1 and 2 and receptors for insulin, IGF-1, and IGF-2 are detectable during the entire preimplantation period, highlighting their importance for growth control. Because all of these growth factors can bind to all the different receptor types (see Figure 8), the different actions are dose and time dependent.

1.2.6 Insulin and fertility

Insulin is an important hormone in the regulation of fertility. The interaction of insulin with other reproductive hormones is necessary because insulin serves as a sensor for metabolic balance, and a certain minimum energy supply needs to be available for reproductive functions such as ovulation (Hill et al., 2008). This is evolutionarily explainable because energy supply must reflect the requirements needed throughout the pregnancy.

Insulin plays a central role in reproduction through direct signalling in the brain. It connects the regulation of the energy status to reproductive functions in the ovary (Bruning et al., 2000). This happens through the hypothalamus–

pituitary–ovarian axis where insulin can influence the gonadotropin-releasing hormone (GnRH) release pattern from hypothalamic neurons in the brain (Sánchez et al., 2012). This interaction is mediated together with growth hormone and IGF. If circulating IGF and insulin levels are low due to maternal malnutrition or undernutrition, basal GnRH secretion is suppressed and/or its pulsatile secretion is depressed (Butler et al., 2003), and this can delay important reproductive events, including the onset of puberty, and can impact on reoccurring events such as ovulation (Garcia-Garcia, 2012).

Importantly, both high-energy conditions as well as undernutrition can have detrimental effects on fertility. Because the ovarian pool is fixed at birth, all oocytes will go through possible periods of metabolic imbalance and possible negative impacts might be retained. This is why the consequences of metabolic stress on fertility might last longer than the stress situation itself. The consequences of cumulus-oocyte complex (COC) or embryo exposure to insulin vary depending on the targeted period of life and the metabolic status of the mother, and this has been shown by several authors. In the study of Armstrong (Armstrong et al., 2001), a diet leading to higher peripheral insulin concentrations was beneficial for the growth of the dominant follicle but impaired oocyte quality. Also, high circulating insulin levels have been shown

to be beneficial for the resumption of cyclicity after parturition (Gong et al., 2002) while this was not valid for heifers with already elevated body weight. A feed regimen designed to lower body weight gain was shown to lead to better embryo yield and quality after collection when compared to heifers with high body weight gain (Freret et al., 2006). Moreover, oocyte quality was reduced in hyperinsulinemic heifers (Adamiak et al., 2005a), and chronic dietary restriction resulted in a reduction in dominant follicle growth rate, diameter, and persistence, and the animals became anoestral (Diskin et al., 2003). These studies highlight the importance of a feeding strategy taking into account the energy requirements for follicular growth without compromising oocyte quality and conception rates (Garnsworthy et al., 2009).

In the ovary, insulin is distributed through the ovarian blood flow and reaches the follicular fluid and thus comes into direct contact with the oocyte via transudation (Poretsky et al., 1999). Insulin receptors are present in all ovarian compartments, such as granulosa and theca cells, as well as in the stroma and on the oocyte (Poretsky et al., 1985), and this shows the importance of insulin signalling on the different ovarian cells. In humans, insulin concentrations in the follicular fluid are positively related to progesterone levels (Diamond et al., 1985) and insulin is thus, as in other mammals, presumed to be involved in oocyte maturation (Totey et al., 1995). In cattle, insulin concentrations in the follicular fluid are dependent on follicular stage, and insulin tends to be higher in preovulatory compared to subordinate follicles. However, the basal level of circulating insulin is influenced by diet, independently of follicle stage (Landau et al., 2000).

Besides a direct role of insulin in follicular development, it can have gonadotropic functions and stimulate steroidogenesis, and the gonadotropic effect of insulin becomes evident in ovarian hypofunction that can be observed in women with type 1 diabetes (Poretsky & Kalin, 1987). Insulin can potentiate the response of ovarian cells to gonadotropins on three levels: 1) through increased LH binding capacity on granulosa cells (Adashi et al., 1985), 2) by sensitizing the pituitary gland to GnRH (Soldani et al., 1994), and 3) by influencing the GnRH secretion of the hypothalamus (Sánchez et al., 2012).

In particular, it is well studied how insulin stimulates steroidogenesis both in vivo and in vitro. Most of the functions of insulin on the ovary are closely linked to the IGF system, including the locally produced IGF binding-proteins (Spicer

& Echternkamp, 1995). Hyperandrogenism in the insulin-resistant state is probably partly transmitted through the IGF receptor that binds insulin if insulin is applied in supra-physiological doses (Poretsky, 1991; Monget & Bondy, 2000). Thus, diet-induced hyperinsulinemia could lead to dysregulation of insulin’s physiological ovarian functions as a mitogen and its effects on the

stimulation of progesterone production in granulosa cells, androgen production in theca cells and progesterone production in luteal cells. (Spicer &

Echternkamp, 1995).

The role of insulin in the final stage of oocyte maturation is based on several facts. Insulin signalling is dependent on the responsiveness of the target tissue through increased expression of insulin receptors. In the preovulatory follicle, insulin receptors are upregulated, and this is associated with increased estradiol production, which supports insulin’s action in final maturation and ovulation.

The increased expression of receptors in the preovulatory period could thus be a crucial factor in transitioning the follicular development to the preovulatory stage (Gong et al., 1991; Shimizu et al., 2008). In humans, insulin acts as a local factor through the insulin receptor, initially appearing in granulosa cells of the preantral follicle (Samoto et al., 1993).

Circulating insulin levels vary during the oestrous cycle with a peak on the day of ovulation, and feeding a diet that leads to hyperinsulinemia significantly reduces the steady-state distribution of insulin receptors in healthy follicles (Armstrong et al., 2001).

1.2.7 Insulin and in vitro embryo production

Insulin signalling components are expressed in cumulus cells (CCs) and oocytes (Acevedo et al., 2007), and insulin signalling is highly important in the very first stages of embryonic development. Transcripts for the insulin and IGF-1 and 2 receptors are detectable at all stages of cattle embryo development from zygote to the blastocyst, which shows the potential of insulin to act at all of these stages (Schultz et al., 1992).

Known for its mitogenic effect, insulin has been used for many years as a stimulatory factor in in vitro cell culture systems and thus also for in vitro embryo production (IVP). In vitro concentrations used in the different media are often much higher (1 to 10 mg/ml) compared with those found physiologically in follicular fluid (0.1 to 1 ng/ml) (Laskowski et al., 2016b), and this is due to the different stabilities of insulin in vitro and in vivo (Hayashi et al., 1978). The reported effects of insulin on early embryonic development vary in the different in vitro studies, and often no effects on the blastocyst rate have been observed (Zhang et al., 1991; Shamsuddin et al., 1993; Bowles & Lishman, 1998; Fouladi- Nashta & Campbell, 2006). Possible positive mechanisms of insulin during embryonic development include anti-apoptotic and mitogenic functions (Byrne et al., 2002). However, elevated insulin might increase oxidative stress through its metabolic function in the cells and thus lower embryo survival, and several reported changes in phenotype or gene expression due to insulin-dependent

signalling illustrates the potential role of insulin acting at several different levels during final oocyte maturation and early embryo development.

1.2.8 Comparative aspects

Situations with elevated or decreased levels of circulating insulin exist in both cattle and humans and in both species, the important role of insulin signalling for maintenance of reproductive functions and the link between metabolism and fertility has been described (Schneider, 2004).

The NEB of the dairy cow postpartum is a well-known example where energy deficiency leads to a failure to maintain reproductive functions and shows the important role of insulin in governing fertility functions (Wathes et al., 2003;

Pryce et al., 2004; Butler, 2005). The consequences of decreased circulating insulin levels for fertility are late resumption of cyclicity postpartum and delayed ovulation, both of which are poor conditions for early embryonic development and for the maintenance of the pregnancy. This effect is mainly transmitted through the role of insulin in the brain, where GnRH and LH pulse frequency is reduced if insulin levels are low, leading to decreased oestrogen production. The same is observed in humans with severe chronic energy deficiency, such as in anorectic women who develop amenorrhea (De Souza & Metzger, 1991) and who have reduced fertility outcomes for both natural conception and assisted reproductive technologies (ART) (Veleva et al., 2008). Furthermore, it is well known that both obesity in women and overfeeding in cows are detrimental for fertility. Hyperinsulinemia is associated with impaired oocyte quality in over- conditioned cows (Adamiak et al., 2005b) just as women suffering from metabolic syndrome or type 2 diabetes have decreased oocyte quality (Niu et al., 2014) and impaired reproductive functions (Sakumoto et al., 2010; Pantasri &

Norman, 2014).

The similarities between the human and bovine species provide an excellent opportunity to elucidate the actions of insulin, and new insights into the relation between metabolic and reproductive disorders could help to improve fertility in both species. Because experiments based on human embryos are ethically controversial, the need for alternative methods to gain further knowledge about underlying mechanisms of impaired fertility due to metabolic imbalance in humans becomes evident. The parallels in human and cattle ovarian reserve, follicular dynamics, and embryo metabolism explain the suitability of the bovine model for human embryonic development (Ménézo & Hérubel, 2002; Campbell et al., 2003a).

1.3 Early embryonic development

1.3.1 Oocyte maturation

Follicular growth until final oocyte maturation takes around three months in the cow in vivo (Lussier et al., 1987; Beam & Butler, 1999), with the final maturation occurring inside the follicle after the release of a preovulatory LH peak (Hyttel et al., 1999). This can be simulated for embryo production by a maturation period of 22–24 h in in vitro systems (Ward et al., 2002). Oocyte maturation is the first fundamental step in the development of a healthy offspring and is a complexly and highly regulated process where the interaction of the oocyte and its surrounding CCs prepares the oocyte on the cellular and molecular level for successful fertilization (Richards, 2005). The COC is an interacting tissue environment where the surrounding CCs support the oocyte during growth and final maturation by supplying it with metabolites. Pyruvate derived from CC glucose metabolism is the preferred substrate to provide the oocyte with energy (Sutton-McDowall et al., 2010, see also a more complete review in chapter 1.3.4.).

The cells of the COC are connected via gap junctions and communicate through paracrine signalling, and the oocyte is dependent on the surrounding cells for the completion of meiotic maturation (Matzuk et al., 2002; Hyttel et al.). The oocyte maturation period is known to be especially sensitive for stressors such as metabolic imbalance, temperature changes, and toxic influences (Moor & Crosby, 1985; Combelles et al., 2009). The early events in life are easily disturbed by such disrupters and the so-called metabolic programming occurs peri-conceptionally (Fowden et al., 2006; Martin-Gronert

& Ozanne, 2012) with potential negative effects for the offspring lasting throughout its life and possibly even transmitted to subsequent generations.

In vivo, COCs are exposed to the follicular fluid, the composition of which is closely correlated to the situation in the maternal serum (Spicer & Echternkamp, 1995; Landau et al., 2000). This fact explains the strong link between nutrition, metabolism, and oocyte quality because metabolites and hormones in the circulation will also come in direct contact with the oocyte (Landau et al., 2000;

Leroy et al., 2012). The maternal nutritional state can programme the oocyte’s metabolism at this early stage of development (O’Callaghan & Boland, 1999;

Fleming et al., 2012).

For final maturation and thus being prepared for successful fertilization, the oocyte has to go through several nuclear and cytoplasmic changes (Hyttel et al., 1986; Eppig, 1996; Fulka et al., 1998).

Briefly, cytoplasmic maturation involves a range of metabolic and structural changes, allowing subsequent fertilization, cell cycle progression from meiosis to mitosis, and activation of several pathways for the programming of preimplantation development (Eppig et al., 1994; Trounson et al., 2001).

From having mitochondria distributed in a peripheral pattern, the LH peak induces the formation of mitochondria clusters associated with lipid droplets. At the same time, the formation of the perivitelline space with loss of contact between CCs appears (Kruip et al., 1983).

On the nuclear level, the nuclear envelope ruffles and meiosis resumes, visible by germinal-vesicle breakdown. In the final stage, the polar body is extruded, the mitochondria disperse, and most organelles move to the centre of the oocyte while cortical granules are formed in the periphery (Kruip et al., 1983;

Combelles et al., 2002).

1.3.2 Development until the blastocyst stage

The fertilized oocyte is called a zygote and contains all of the materials for initiating the first developmental steps, including new protein synthesis, mRNA activation, protein and RNA degradation, and reorganization of the organelles in the cell (Stitzel & Seydoux, 2007). The first cleavage occurs 25–26 h post fertilization (Hamilton & Laing, 1946; Sakkas, 2001). The first cell divisions are under maternal control and are based on stored mRNA and protein molecules in the oocyte before the embryonic genome takes over transcription (Barnes &

Eyestone, 1990) (see more in section 1.3.1). At 42–44 h post fertilization, the embryo should have reached the 4- to 8-cell stage with equal numbers of blastomeres (Betteridge & Fléchon, 1988). Early cleavage dynamics have been reported to be a tool to predict embryo quality and developmental potential (Van Soom et al., 1992; Kubisch et al., 1998; Lonergan et al., 1999), and the fastest- growing embryos seem to have the best viability and potential to reach the morula and blastocyst stages.

The next ultrastructural change is morula formation at the 32-cell stage when compaction occurs (Van Soom et al., 1992). The blastomeres become either part of the embryonic inner cell mass (ICM) or the trophoblast that will form the foetal annexes (Betteridge & Fléchon, 1988). In the bovine, the transition from compacted morula to the blastocyst stage occurs between day 6 and 8 of development (Betteridge & Fléchon, 1988). The most important characteristic of a blastocyst is the formation of the blastocoel cavity and the clearly distinguishable ICM. The embryo is enclosed by the zona pellucida until hatching (Lindner & Wright, 1983).

1.3.3 Embryo morphology

The most accurate evidence for the developmental competence of an embryo is to allow development into a live, healthy offspring. This is, for several reasons, not always applicable for research purposes. Thus, other methods have been established with the aim of predicting oocyte developmental potential and embryo quality (Van Soom et al., 2003).

Basic developmental data at different time points are usually recorded to follow the different developmental steps and to look for signs that the embryo might fail to pass the important thresholds such as first cleavage, embryonic genome activation (EGA), compaction, and blastocyst formation (Andra et al., 1999; Lonergan et al., 2006). On Day 7 and 8, morphological evaluation of blastocyst stages and quality grading of blastocysts is possible at a more advanced level and includes several criteria. The diameter of a Day-8-blastocyst (BC8) is approximately 150 to 190 —m, including a zona pellucida thickness of 12 to 15 —m, and some parameters that are included in the evaluation of embryo quality are shape, colour, cell number, presence of extruded and degenerated cells and size of the perivitelline space (Lindner & Wright, 1983, Crosier et al., 2001). In addition to these, staining for the actin skeleton, mitochondrial pattern (Zijlstra et al., 2008), lipid droplets (Abe et al., 1999, 2002), and apoptosis (Yang et al., 1998; Gjørret et al., 2003) can be used to detect differences in embryo phenotype and allows, together with the developmental rates, conclusions to be drawn about the embryo's viability.

The first important assessment criterion is light microscopy determination of the developmental stage (Shea, 1981) to ensure that the embryo is not retarded in development compared to the other embryos of the same day of development.

On Day 8, the blastocyst can be early, blastocyst, expanding, expanded, hatching, or hatched and the different blastocyst stages were described and defined by Lindner and Wright in 1983 (Lindner & Wright, 1983), and the same staging criteria are still used today. In brief, an early blastocyst has already formed a fluid-filled blastocoel and the embryo itself forms around 70 -80% of the volume. The blastocyst stage is characterised by a growing blastocoel that is highly prominent, and compaction of the embryo (differentiation between ICM (darker) and trophoblast) becomes visible. Up to this stage, the diameter of the embryo does not change much from the size of the oocyte. When the embryo is expanding or expanded, the zona pellucida thins and the embryo diameter increases. The next stage is the partly hatched (hatching) or entirely hatched embryo, at which point it has shed the zona pellucida. All of these stages can be found on Day 8 of development.

The next step is to assess the embryo quality grade according to the guidelines that have been developed by the International Embryo Technology

Society (IETS) (Stringfellow DA, 2010). The quality grades are indicated by a descending scale between 1 and 4 where 1 stands for “excellent/good”, 2 for

“fair”, 3 for “poor” and 4 for “dead/degenerated”.

Following different types of staining, more characteristics can be evaluated such as cell number, mitochondria distribution, and actin cytoskeleton structure.

A blastocyst on Day 8 contains around 100–200 cells (Byrne et al., 1999;

Watson et al., 2000), and larger embryos are often assessed as more viable because they are further advanced in their development. However, there are also other theories that claim that moderate growth is beneficial for the long-term viability of the embryo and health of the offspring (Leese et al., 2008).

Mitochondria are the energy providing organelles in cells and are reported to have important functions in competent oocytes and blastocysts (Lane & Gardner, 1998; Bavister & Squirrell, 2000). The mitochondrial activity and pattern in the embryo varies depending on developmental stage (Tarazona et al., 2006).

However, an even distribution between all blastomeres with no accumulations or empty areas within the blastocyst is considered to be beneficial for the viability of the embryo (Båge et al., 2003; González & Sjunnesson, 2013). The same is true for actin distribution in all parts of the embryos where good quality embryos often show an equal distribution pattern, while degraded or low quality embryos show cytoskeleton disintegration and might thus be more fragile and less viable (Zijlstra et al., 2008; González & Sjunnesson, 2013).

1.3.4 Oocyte and embryo metabolism

Oocyte maturation and the first developmental steps such as growth, cell division, and differentiation are energy-consuming processes that require a high availability of energy substrates (Gardner, 1998). The embryo itself or through the CCs is able to metabolize different types of substrates such as glucose, triacylglycerides, and amino acids. Besides these exogenous substrates that are present in oviductal fluid or in vitro media, endogenously stored triacylglycerides and glycogen are reported to contribute to energy availability during early development (Brinster, 1971; Ferguson & Leese, 2006). Before elongation, embryos are reliant on oxidative phosphorylation of pyruvate, lactate, and amino acids for ATP production, with increasing glucose consumption in more advanced stages after compaction (Leese, 1995;

Thompson et al., 1996; Thompson, 2000). Besides providing energy, carbohydrate and amino acid metabolism generates substrates with functions in the cellular stress response (Gardner, 1998).

Oocyte metabolism is tightly connected to CC metabolism (Figure 4), and in the early stages of development the CCs provide the oocyte with pyruvate, its

most favourable substrate (Sutton-McDowall et al., 2010). Oocyte-CC communication continues even after meiotic resumption when most of the gap junctions are lost (Sutton et al., 2003). Oxygen consumption is a good measure of metabolic activity, and during maturation it is at similar levels as at the blastocyst stage (Houghton & Leese, 2004) and the total ATP content is increased in mature oocytes (Stojkovic et al., 2001; Ferguson & Leese, 2006). It has been shown that LH increases glycolysis in bovine oocytes, and this is assumed to be the mechanism through which LH enhances maturation (Zuelke

& Brackett, 1992).

Glucose consumption has been used to predict embryos with high developmental potential in mice (Gardner & Leese, 1987) and cattle (Renard et al., 1980). However, moderate glycolytic activity close to levels that are found in vivo seems to be best for viability of the embryo (Lane and Gardner, 1996).

Here, elevated insulin levels could contribute to excess glucose consumption during early embryo development and thus have an adverse effect on viability.

Glucose metabolized through the pentose phosphate pathway generates ribose that can be used for nucleic acid production and NADPH for glutathione regeneration through reduction, an important pathway against ROS (Wales, 1973; Rieger, 1992; Gardner, 1998).

Lipid metabolism also varies depending on development stage and can be of exogenous or endogenous sources (Ferguson & Leese, 2006; Haggarty et al., 2006). In humans, it has been shown that pre-implantation embryos actively take up fatty acids (Haggarty et al., 2006). Besides triacylglycerides and fatty acids, even cholesterol has important functions during embryogenesis, and impaired cholesterol metabolism might have detrimental consequences for the embryo (Farese & Herz, 1998).

Amino acids are possibly consumed to a different extent depending on embryo developmental stage (Partridge & Leese, 1996; Lane & Gardner, 1998).

Alanine and glutamine are precursors for gluconeogenesis (Felig et al., 1970), but their contribution to energy supply is limited. Other functions of amino acids might be of higher relevance during embryo development because they function as substrates for biologically important molecules such as melanin (Korner &

Pawelek, 1982), as osmolytes, as buffers, and as regulators of embryo metabolism (reviewed by Gardner, 1998), and this explains their role in embryo development (Takahashi & First, 1992).

In summary, any dysregulation in substrate availability or metabolic functions might lead to disturbances in healthy embryo development. This highlights the importance of developing adequate media for IVP of embryos that is similar to the conditions found in vivo because requirements for the different metabolites could change during development, and it also explains why

metabolic disturbances in the mother might lead to conditions that are suboptimal for embryo viability.

Figure 4. Proposed model of the metabolic interactions and activity of CCs and the oocyte.

Numerous energy substrates are supplied to the COC by the surrounding fluid, including glucose, pyruvate, lactate, and amino acids.

Glucose can be utilized via three major pathways: (i) glucose oxidation (the combination of glycolysis, tricarboxylic acid (TCA) cycle and oxidative phosphorylation); (ii) the pentose phosphate pathway (PPP); or (iii) it can be converted to intermediates and utilized for extracellular matrix (ECM) expansion. FSH stimulates glucose metabolism by cumulus cells. Glucose utilization begins with glycolysis (within cumulus cells) where glucose-6-phosphate is converted to pyruvate, which can then enter the oocyte directly or be converted to lactate.

Pyruvate is further oxidized by the TCA cycle within ovum mitochondria, followed by oxidative phosphorylation in the mitochondrial intermembrane where ATP is released by electron transfer. PPP also begins with the oxidation of glucose to glucose-6-phosphate within cumulus cells, with one of the products of the pathway, phosphoribosyl pyrophosphate (PRPP), being used by the oocyte for purine synthesis. Purines are involved in the regulation of nuclear maturation. PPP is also involved in general cytoplasmic homeostasis since NADP+ is reduced to NADPH. Amino acids cystine and cysteine are involved in the production of glutathione (GSH), accumulation of which appears essential for early embryonic development. Although oocyte-secreted factors are known to have major effects on development and differentiation of cumulus cells, there are no data available concerning their effects on the metabolism of cumulus cells. GSSG= oxidized GSH; ROS = reactive oxygen species. Reprinted with permission from (Sutton et al., 2003).

1.3.5 In vitro produced embryos

Compared to their in vivo counterparts, IVP embryos are in general less viable.

Blastocyst development rates are lower in vitro, and differences in morphology, metabolism, gene expression, epigenetics, cryotolerance and pregnancy rates after transfer have been studied in order to explain their decreased developmental competence (Wright & Ellington, 1995; Thompson, 1997).

One explanation for these differences is an increased exposure to oxidative stress during in vitro culture (Cagnone & Sirard, 2013; de Assis et al., 2015).

Many attempts to improve culture conditions and protocols have been made in recent decades resulting in improved IVP, but differences in embryo quality are still observed (Niemann & Wrenzycki, 2000; Galli et al., 2003; Hasler, 2003).

For example, Khurana and Niemann (2000b) detected differences in aerobic glycolysis rate and lactate oxidation between IVP and in vivo embryos.

Some authors observed differences in morphology in the form of a darker overall appearance with larger blastomeres at early stages and a reduced perivitelline space along with reduced viability of IVP, with fewer IVP embryos surviving the cryopreservation procedures (Khurana & Niemann, 2000a; Rizos et al., 2002). Even if not having a consistently different morphology, a mismatch in timing might exist in IVP embryos that leads to delays in development (Plante

& King, 1994).

In vitro, typical developmental rates for bovine embryos are 20–40%

(Thompson & Duganzich, 1996; Rizos et al., 2002), and gene expression studies have been used with the aim of discovering the underlying reasons for the impaired developmental potential of IVP embryos. Important genes for development are differentially expressed in IVP embryos, e.g. connexion 43, which plays a role in maintaining compaction (Wrenzycki et al., 1996; Niemann

& Wrenzycki, 2000). Also, CCs have a different gene expression pattern depending on whether oocyte maturation occurs in vitro or in vivo (Tesfaye et al., 2009; Gad et al., 2012)

Current research efforts into gene expression and regulation aim to better understand the epigenetic mechanisms that present a link between the genome and the environment and induce permanent changes in gene expression pattern through metabolic programming (Santos et al., 2003) (see chapter 1.4.2.)

1.3.6 Comparative aspects between human and bovine embryo development

The use of animal models to better understand mammalian embryogenesis is important because such models allow the study of pathophysiological mechanisms in the oocyte and early embryo without evoking ethical